With nearly 300 scientific citations, Oncomine has become an industry standard for cancer researchers. This series showcases recent scientific publications that use Oncomine in various and novel ways to further cancer research.

03 FEB 2010
Inactivation of clusterin contributes to tumorigenesis in prostate cancer mouse models.

Researchers from the University of Parma & Modena, Institute of Child Health in London, the Thomas Jefferson University, and MedImmune recently reported that the loss of clusterin (CLU) expression resulted in prostate epithelial transformation, metastatic spread, and tumorigenesis¹. These authors provide a mechanistic model for these events by demonstrating an increase in NFKB expression in CLU knockout mice. To understand the functional role of CLU in prostate cancer the team generated CLU mouse knock-outs and made the following observations:

• transformation of prostate epithelium to PIN or differentiated carcinoma in 100 percent of mice with homozygous deletions in CLU, and in 87 percent of heterozygotes
• enhanced metastatic spread in CLU mouse knock-outs crossed with prostate cancer prone (TRAMP) mice
• tumorigenesis in female CLU knock-out TRAMP mice normally free of disease
• increase NF-KB activity in CLU knock-outs

To determine the relevance of this mouse data for understanding human disease, study authors used Oncomine to evaluate CLU expression across a wide range of human primary tumor samples. Consistent with the mouse data, an Oncomine gene search on CLU shows significant under-expression of CLU in a number of Prostate Cancer vs. Normal analyses, and in Metastatic vs. Primary and Pathology Grade analyses (Fig. 1). This Disease Summary view also indicates that several other cancer types have significant under-expression results for CLU in Cancer vs. Normal, Pathology, Grade, and Stage analyses. Each of these offer additional opportunities for further exploration.

Figure 1: Disease Summary for CLU - This view displays the number of significant results colored red or blue for over- or under-expression, respectively, across all cancer types and analysis types in Oncomine. This high level summary can be used to drill down on multiple results within a cell by clicking on that cell and browsing the supporting evidence. Here there are five under-expression results (blue) for Prostate Cancer vs. Normal and one for Metastasis and Stage analyses at the default statistical threshold of Top 1% genes under-expressed and having p-value of 1E-4 or better.

Focusing on Prostate Cancer vs. Normal analyses, a number of datasets generated by independent investigators confirm the significant under-expression of CLU in prostate cancer relative to normal prostate gland (Fig. 2). These results provide strong supporting evidence that CLU mRNA is indeed under-expressed in prostate cancer in human disease.

Figure 2: Significant under-expression of CLU in prostate cancer relative to normal prostate was observed in several Oncomine studies. This figure shows results from the Welsh², LaPointe³, Liu⁴, and Vanaja⁵ studies where each bar represents the measurement of CLU in an individual sample, with normal samples grouped in Class 1 and prostate cancer samples grouped in Class 2.

Data in Oncomine also supports a model in which under-expression of CLU is associated with prostate metastasis in human disease, consistent with findings from the mouse models (Fig. 3).

Figure 3: Significant under-expression of CLU in prostate cancer metastases is validated across multiple independent studies.

Extending this analysis further, we utilized the sophisticated functionality built into Oncomine’s user interface to query a meta-analysis of five studies with low CLU expression with a list of genes known to be directly targeted by NF-KappaB - and potentially up-regulated when CLU is under-expressed. To accomplish this we layered the following information to produce a single meta-analysis comparison map:

1. Oncomine gene search on CLU
2. Oncomine Cancer Type search on Prostate
3. Oncomine Analysis Type search on Cancer vs. Normal
4. Oncomine Concept search on “NF-KappaB”, followed by selection of “Concept: NF-kappaB (B) - Transfac Transcription Factor Targets” from the auto prompt (1)
5. Selection of analyses for comparison using check boxes (2)
6. Selection of Under-expression gene rank from ORDER BY menu
7. Selection of Compare to build a comparison map (3)
8. Sort by over-expression in the meta-analysis view (4) (Fig. 4)

Figure 4: After identifying under-expression of CLU in Prostate Cancer vs. Normal analyses, here we add a Concept (NF-KappaB target genes) as a filter, select analyses to compare and visualize a ranked list of over-expressed genes in the concept that validate across the datasets.

Note that by selecting the under-expression gene rank for CLU in the list selector, and the over-expression gene list from the drop-down menu on the right, we generate a list of NF-KappaB target genes that are over-expressed in Prostate Cancer vs. Normal analyses in which CLU is significantly under-expressed (Fig 5).

Figure 5: Comparison map of NF-KappaB transcription factor target genes that are over-expressed in multiple independent datasets where CLU is under-expressed.

The result of this query is the set of known NFKB target genes that are over-expressed across five independent Prostate Cancer vs. Normal analyses in which CLU was under-expressed. The first gene on the list, hepsin (HPN), is a known prostate cancer marker. This ranked list of genes can be further evaluated either individually or by uploading the list as an Oncomine Concept to identify association of the targets with other prostate cancer studies.

This study leverages the wealth of Oncomine primary tumor data to support hypotheses that further the understanding of the role of CLU and its potential relationship with NF-KappaB activation. Oncomine provides the data and analysis tools required to dissect the complexity of signaling events in cancer by identifying underlying genes involved in this process, and to validating them across a range of datasets and analysis types. Further investigation of the resulting list of genes can be undertaken through gene searches and signature mapping in Oncomine to better understand their individual and collective role in prostate cancer.

¹ Genetic inactivation of ApoJ/clusterin: effects on prostate tumourigenesis and metastatic spread.
Bettuzzi S, Davalli P, Davoli S, Chayka O, Rizzi F, Belloni L, Pellacani D, Fregni G, Astancolle S, Fassan M, Corti A, Baffa R, Sala A.
Oncogene. 2009 Dec 10;28(49):4344-52.

² Analysis of gene expression identifies candidate markers and pharmacological targets in prostate cancer.
Welsh JB, Sapinoso LM, Su AI, Kern SG, Wang-Rodriguez J, Moskaluk CA, Frierson HF Jr, Hampton GM.
Cancer Res. 2001 Aug 15;61(16):5974-8.

³ Gene expression profiling identifies clinically relevant subtypes of prostate cancer.
Lapointe J, Li C, Higgins JP, van de Rijn M, Bair E, Montgomery K, Ferrari M, Egevad L, Rayford W, Bergerheim U, Ekman P, DeMarzo AM, Tibshirani R, Botstein D, Brown PO, Brooks JD, Pollack JR.
Proc Natl Acad Sci U S A. 2004 Jan 20;101(3):811-6. Epub 2004 Jan 7.

⁴ Sex-determining region Y box 4 is a transforming oncogene in human prostate cancer cells.
Liu P, Ramachandran S, Ali Seyed M, Scharer CD, Laycock N, Dalton WB, Williams H, Karanam S, Datta MW, Jaye DL, Moreno CS.
Cancer Res. 2006 Apr 15;66(8):4011-9.

⁵ Transcriptional silencing of zinc finger protein 185 identified by expression profiling is associated with prostate cancer progression.
Vanaja DK, Cheville JC, Iturria SJ, Young CY.
Cancer Res. 2003 Jul 15;63(14):3877-82.

Oncomine 4.2.3 | Data (September 2009)